Soil moisture assessment and modeling can be defined as one of the indicators of smart agriculture. Soil moisture greatly affects plant growth and, consequently, agricultural productivity. Field methods for estimating soil moisture, from θm to TDR, are highly accurate at the local level, but they are very costly and time-consuming for a large area. Remote sensing models and technology can be used to reduce costs and time and provide global estimates of soil moisture. In the present study, by using various remote sensing indices obtained from different bands of Landsat 8-9 and Sentinel 2 satellites, spatio-temporal modeling of soil moisture was performed using stepwise regression, geographical weighted regression, and random forest models. According to the results, the NSMI index was one of the effective indices with positive correlation, with a Pearson coefficient of more than 75 percent and a RMSE of 0.018. Also, scenarios 3 and 4 had the best correlation with surface soil moisture, and the correlation between these two scenarios provided accurate estimates of soil moisture with an RMSE range of 0.011-0.026. Therefore, the use of different satellite spectral bands is recommended as an indirect, fast, low-cost, and acceptable-accuracy method for estimating surface soil moisture and regular monitoring of this important hydrological and climatic variable.